[coreboot] Asus E35M1-M PRO Fusion AMD E-350 APU

Florian Zumbiehl florz at florz.de
Mon Jul 25 03:52:53 CEST 2011


Hi,

> > but so far memory initialization fails most
> > of the time (as in: I had two or three boots that reached the payload
> > so far). The (apparently) same failure has been observed with the ASRock
> > board as well:
> > http://www.coreboot.org/pipermail/coreboot/2011-July/065795.html
> > 
> > As such, I am actually currently not really working on it, but rather
> > waiting for Frank Vibrans to have a closer look at the AGESA debug output
> > the board generates before it stops.
> 
> I did not see any message on this list regarding your problem. Could you
> open a new thread so that people searching the Web know what the current
> state is and maybe somebody else reading the list has an idea.

That all happened on IRC, and I guess there isn't really much interesting
to report on so far, other than what has already been said. I do have a
~ 250 kB debug log, if anyone is interested, but because of its size I
wouldn't want to just broadcast it over the list.

> > > bios is UEFI based though.
> > 
> > ... and horribly broken, yes.
> 
> Can you elaborate in what regard please to give us some arguments when
> talking about coreboot.

Well, it has a graphical setup program with mouse support and an "easy
mode". Do I need to say more? Well, maybe this: It uses a 640x480 video
mode on a monitor that would do 1024x768 just fine, but obviously the
interface has been designed for 1024x768 only, so it cuts off lots
of stuff and is essentially unusable with that monitor. Well, except if
you give it a full HD monitor first to read EDID from, then it produces
output that looks horrible when scaled by that monitor, but works just fine
with the old 1024x768 CRT.

> > > Any thoughts/idea's?
> > 
> > Well, I can provide you with what's necessary to get the serial port going,
> > but currently there isn't much to do unless you are familiar with
> > AMD CPU and chipset initialization, I guess.
> 
> If you have the means to publish your current state as a Git repository
> somewhere that would be appreciated I guess.
> 
> Additionally I do not know what the best way is to send patches to
> support a new board. One huge patch which gets the board going right
> away or several small patches like serial port, memory initialization, ???
> and the last one being adding it to the Konfig system.

We will see--so far, it's just some minimal superio initilization to get
the serial port going, plus a few changes to the board config (based on the
ASRock one, that is) that are essentially untested due to the nature of the
current obstacle, and which probably are incomplete anyway.

If anyone is seriously interested in working on this, let me know, but
other than that I'd first try to get this to a state where there are any
real patches to speak of before I put any effort into publishing things :-)

Florian




More information about the coreboot mailing list